Sarti Alessandro, Citti Giovanna
CNRS-EHESS, Paris, France,
J Comput Neurosci. 2015 Apr;38(2):285-300. doi: 10.1007/s10827-014-0540-6. Epub 2014 Dec 23.
In this paper we show that the emergence of perceptual units in V1 can be explained in terms of a physical mechanism of simmetry breaking of the mean field neural equation. We consider a mean field neural model which takes into account the functional architecture of the visual cortex modeled as a group of rotations and translations equipped with a degenerate metric. The model generalizes well known results of Bressloff and Cowan which, in absence of input, accounts for hallucination patterns. The main result of our study consists in showing that in presence of a visual input, the stable eigenmodes of the linearized operator represent perceptual units of the visual stimulus. The result is strictly related to dimensionality reduction and clustering problems.
在本文中，我们表明初级视皮层（V1）中感知单元的出现可以用平均场神经方程的对称性破缺这一物理机制来解释。我们考虑一个平均场神经模型，该模型考虑了视觉皮层的功能结构，将其建模为一组配备退化度量的旋转和平移。该模型推广了布雷斯洛夫和考恩的著名结果，即在没有输入的情况下，该结果解释了幻觉模式。我们研究的主要结果在于表明，在存在视觉输入的情况下，线性化算子的稳定本征模代表视觉刺激的感知单元。该结果与降维和聚类问题密切相关。
